Grid Computing at Scale for Financial Services
Article Synopsis :
This whitepaper from Amazon Web Services explores the massive potential of grid computing in Financial Services (FS) including known challenges and potential use cases in risk management, product development, quantitative research, and capital management.
Heightened regulatory requirements and dynamic market conditions are forcing companies to consider expanding their on-premise grid-computing capabilities – which can be capital-, labor-, and time-intensive. Hence, grid-computing options in the cloud – from vendors such as AWS, Microsoft and Google – are growing in popularity.
Financial simulations are vital to helping FS institutions identify and manage risk, fully comprehend capital positions, and make informed investment and pricing decisions, among other uses. What has changed to make such simulations more onerous? Three things:
- Regulatory bodies are requiring FS institutions to perform increasingly stringent stress tests to maintain adequate capital ratios.
- Obligatory solvency requirements are further complicated by stipulations related to capital and collateral to maintain specified liquidity levels in order to conduct business.
- The development of new quantitative trading strategies and complex products requires a greater variety of data sets and increased modelling capacity, which in turn adds to the complexity of design and back testing.
These new burdens, which are exacerbated by limited datacenter capacity, affect organizations across the FS industry and can lead to significant delays in job completion. These delays can in turn deter users from running jobs altogether, which introduces a range of financial and operational risks.
The challenge of building and maintaining an effective on-premise grid boils down to the difficulty of meeting variable demand with a fixed resource. It’s problematic to correctly size a static grid that can effectively accommodate shifting daily computing needs from times of low usage to peak demand. Particular issues impacting the efficacy of on-premise grids include:
- The expanding volume of data financial institutions need to gather, process, analyze, and store to run meaningful, accurate calculations
- On-premise hardware flexibility is limited: since not all servers are optimized for grid-computing workloads, processing times often lag
- If markets experience volatility, regulations change, or a team wants to experiment with a new product, grid requirements often spike
Faced with the limitations of on-premise storage and computing capabilities, FS institutions of all sizes—fintech startups, hedge funds, large insurers, and global investment banks, among others—are working with cloud vendors to increase their grid-computing capabilities by extending their on-premise grids to the cloud or by building cloud-native grids. There are seven advantages to this approach:
- Virtually unlimited compute and storage resources
- Various instance types for specific types of workloads
- Cost optimization with a variety of pricing structures
- Enhanced (and automated) security and compliance
- Expanded big data capabilities for analysis and business intelligence
- Automation capabilities for scaling and provisioning resources
- Management consoles to provide enterprise-wide control and visibility
Common use cases driving adoption of cloud-enabled grid-computing capabilities include:
Capital management and reporting: To compel FS organizations to adopt more efficient capital measures, regulations such as CCAR, Solvency II, and the upcoming FRTB are requiring formal reporting around an organization’s ability to undergo and withstand capital stress testing. These requirements have increased the demand and frequency for simulation-based measures for market risk, credit, and counterparty.
Product and strategy development: The development of new financial products requires extensive historical back testing and market simulations.
Risk management and portfolio optimization: Portfolio simulations and scenario testing allow portfolio managers to (1) identify potential risks within their portfolio of products, highlighting hedging and portfolio optimization opportunities, and (2) model the impact of hypothetical portfolio changes.
Contract pricing and valuation: The industry relies heavily on simulations for the pricing and valuation of financial products, including, for instance, credit and interest rate derivatives and variable annuities using stochastic models with heavy use of Monte-Carlo simulation methods.
Link to Full Article:: click here
Digital Insurer's CommentsGrid computing systems utilize standard, open, general-purpose protocols and interfaces to link remote hardware to perform very large tasks. Think: virtually infinite compute power.
Once the exclusive domain of large-scale mostly theoretical R&D and Engineering and Design, grid computing is gaining traction in the FS enterprise for business analytics and product development. Cloud vendors now put grid computing in the hands of anyone with a credit card and start-ups – including InsurTech – are flocking to the technology.
Link to Source:: click here